• Rashi Arora, James P Campbell, Gyorgy Simon, and Nishant Sahni.
• Division of General Internal Medicine, Department of Medicine, University of Minnesota, 420 Delaware Street, SE, MMC 741, Minneapolis, MN, 55455, USA.
• Infection. 2017 Jun 1; 45 (3): 291-298.

BackgroundPhysicians frequently rely on the systemic inflammatory response syndrome (SIRS) criteria to detect bloodstream infections (BSIs). We evaluated the diagnostic performance of procalcitonin (PCT) in detecting BSI in patients with and without SIRS.MethodsWe tested the association between BSI, serum PCT levels, contemporaneous SIRS scores and serum lactate using logistic regression in a dataset of 4279 patients. The diagnostic performance of these variables was assessed.ResultsIn multivariate regression analysis, only log(PCT) was independently associated with BSI (p < 0.05). The mean area under the curve (AUC) of PCT in detecting BSI (0.683; 95% CI 0.65-0.71) was significantly higher than serum lactate (0.615; 95% CI 0.58-0.64) and the SIRS score (0.562; 95% CI 0.53-0.58). The AUC of PCT did not differ significantly by SIRS status. PCT of less than 0.1 ng/mL had a negative predictive value (NPV) of 97.4 and NPV of 96.2% for BSI in the SIRS-negative and SIRS-positive patients, respectively. A PCT of greater than 10 ng/mL had a LR of 6.22 for BSI in SIRS-negative patients. The probability of BSI increased exponentially with rising PCT levels regardless of SIRS status.ConclusionThe performance of PCT for the diagnosis of BSI was not affected by SIRS status. Only PCT was independently associated with BSI, while the SIRS criterion and serum lactate were not. A low PCT value may be used to identify patients at a low risk for having BSI in both settings. An elevated PCT value even in a SIRS negative patient should prompt a careful search for BSI.

22 keywords...

hide…